> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mov-eat.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Coaching context

> How Moveat should store personal coaching information without polluting session or account responses.

# Coaching context

Coaching context is long-lived user information that helps Moveat personalize advice.

It is different from account data, profile data and conversation state.

## Purpose

Coaching context helps answer questions such as:

* What motivates this user?
* What usually blocks adherence?
* What tone works better for them?
* What fears or concerns do they have?
* What constraints affect their plan?

## Why it should be separate

The `/me` session response should stay small and focused on identity/session state. Coaching context can become large and sensitive, so it should live behind dedicated APIs.

## Product examples

Good coaching context examples:

* "Prefers direct feedback and short explanations."
* "Usually struggles with late-night snacking."
* "Motivated by strength progress more than scale weight."
* "Has limited time during weekdays."

Bad coaching context examples:

* Raw chat transcripts.
* Unverified medical claims.
* Secrets or credentials.
* Data that belongs in structured profile fields.

## Agent usage

Agent can request coaching context from Platform and cache it briefly in Redis.

Redis should not be the only durable storage for this data. Platform remains the source of truth.